Glossary
Last-click attribution: the model most people use by accident
Last click gives the whole sale to the final ad. What it hides, when it still works fine, and what to fix before switching to another model.
Also available in: Português · Español
Last-click attribution gives the entire conversion credit to the final clicked interaction before it. It is the default in most tools and, in practice, the model almost everyone uses without ever having chosen it.
How it works
Someone sees an Instagram ad on Monday, searches your brand on Google on Wednesday, clicks the brand ad and buys. Under last click, the sale belongs to Google Ads. Instagram, which created the search, gets zero.
None of that is technically wrong. The model does exactly what it promises. The problem is the decision that follows: someone reads the report, sees Google performing and Meta not, and moves budget from the channel that created demand to the channel that harvested it.
What it hides
- The discovery channel. Whoever introduces the brand is almost never whoever closes.
- The brand campaign. It looks brilliant in any account, because it harvests intent created elsewhere.
- The shutdown effect. Pause the top-of-funnel channel and the brand campaign empties two weeks later. That is the simplest test of a dependency the model will never show.
| Situation | Does last click work? |
|---|---|
| Short cycle, single channel | Yes, with no complication |
| Impulse purchase, low ticket | Yes, with a caveat on remarketing |
| Long journey across channels | No, it underrates the top |
| Account running Meta and Google together | No, it always credits search |
When it still works
The model deserves a defense: it is simple, stable and auditable. Nobody argues about where the number came from. In single-channel operations with short cycles and small budgets, switching models adds complexity without improving a single decision.
The mistake is not using last click. It is using it without knowing you are, and then treating the report as if it described cause.
What to fix before switching models
Changing the model inside the platform solves less than it seems, because each platform still sees only its own traffic. What works, in order:
- Record the source at signup. One source field filled in the CRM beats any model. The standard is in UTM parameters for ads.
- Compare sales by source, not conversions by platform. It is the only ruler where Meta and Google fit together.
- Test by turning things off. Cutting one channel for two weeks and watching the other says more about cause than any model.
The crediting periods behind all of this are in attribution window, and the broader scoreboard in marketing KPIs.
The format that holds both readings together is on the paid media reporting page.
Frequently asked questions
What is last-click attribution?
The model that gives one hundred percent of the conversion credit to the last clicked interaction before the sale. Everything earlier in the journey gets zero.
Is last click bad?
It is not bad, it is partial. It is simple, auditable and fine for short cycles with one dominant channel. It misleads when the journey crosses several channels, because it always rewards whoever closes and never whoever opened.
What should I use instead?
Before switching models, record the source at signup and compare sales by source in the CRM. That fixes more cases than changing the model inside the platform.